New models listed as supported in the docs, like Qwen 3.6 35B A3B or Kimi K2.6, don't seem to work

[lebovic]

I'd like to try Tinker for additional RL / LoRA on Qwen 3.6 35B A3B and Kimi K2.6. Both are listed in the docs, but they don't seem to work for initial sampling.

Are these models supported yet?

---

K2.5 works but K2.6 fails

curl -sS https://tinker.thinkingmachines.dev/services/tinker-prod/oai/api/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $TINKER_API_KEY" \
  -d '{"model":"moonshotai/Kimi-K2.5","messages":[{"role":"user","content":"Reply with Hello."}],"max_tokens":4}'

200 with "Hello"

curl -sS https://tinker.thinkingmachines.dev/services/tinker-prod/oai/api/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $TINKER_API_KEY" \
  -d '{"model":"moonshotai/Kimi-K2.6","messages":[{"role":"user","content":"Reply with Hello."}],"max_tokens":4}'

400 with "Sampling is not supported for moonshotai/Kimi-K2.6."

Qwen 3.5 35B A3B works but 3.6 fails

curl -sS https://tinker.thinkingmachines.dev/services/tinker-prod/oai/api/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $TINKER_API_KEY" \
  -d '{"model":"Qwen/Qwen3.5-35B-A3B","messages":[{"role":"user","content":"Reply with Hello."}],"max_tokens":4}'

200 with "Hello"

curl -sS https://tinker.thinkingmachines.dev/services/tinker-prod/oai/api/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer $TINKER_API_KEY" \
  -d '{"model":"Qwen/Qwen3.6-35B-A3B","messages":[{"role":"user","content":"Reply with Hello."}],"max_tokens":4}'

400 with "Sampling is not supported for Qwen/Qwen3.6-35B-A3B."

---

[dphuang2]

you caught us a little early! We are in the process of getting these live on Tinker, please standby 🙏